]> git.p6c8.net - dumbdbm.git/blob - DumbDBM_File.pm
Fixed some things in the POD documentation
[dumbdbm.git] / DumbDBM_File.pm
1 package DumbDBM_File;
2
3 # DumbDBM_File - Portable DBM implementation
4 #
5 # Based on Python's dumbdbm / dbm.dumb
6 #
7 # Author: Patrick Canterino <patrick@patrick-canterino.de>
8 # License: 2-Clause BSD License
9
10 use strict;
11
12 use Carp qw(carp croak);
13 use Fcntl;
14
15 our $VERSION = '0.1';
16
17 our $_BLOCKSIZE = 512;
18
19 sub _update {
20 my $self = shift;
21 local *FILE;
22
23 $self->{'_index'} = { };
24
25 open(FILE,'<'.$self->{'_dirfile'}) or carp $!;
26
27 while(<FILE>) {
28 my $line = $_;
29 $line =~ s/\s+$//g;
30
31 my ($key,@pos_and_siz_pair) = eval($line);
32 $self->{'_index'}->{$key} = \@pos_and_siz_pair;
33 }
34 }
35
36 sub _commit {
37 my $self = shift;
38
39 unlink($self->{'_bakfile'});
40 rename($self->{'_dirfile'},$self->{'_bakfile'});
41
42 open(FILE,'>'.$self->{'_dirfile'}) or carp $!;
43
44 while(my($key,$pos_and_siz_pair) = each(%{$self->{'_index'}})) {
45 print FILE "'$key', ($pos_and_siz_pair->[0], $pos_and_siz_pair->[1])\n";
46 }
47
48 close(FILE);
49 }
50
51 sub _addval {
52 my ($self,$val) = @_;
53 local *FILE;
54
55 open(FILE,'+<'.$self->{'_datfile'}) or carp $!;
56 binmode(FILE);
57 seek(FILE,0,2);
58
59 my $pos = tell(FILE);
60 my $npos = int(($pos + $_BLOCKSIZE - 1) / $_BLOCKSIZE) * $_BLOCKSIZE;
61
62 print FILE "\0" x ($npos-$pos);
63
64 $pos = $npos;
65
66 print FILE $val;
67
68 close(FILE);
69
70 return ($pos,length($val));
71 }
72
73 sub _setval {
74 my ($self,$pos,$val) = @_;
75 local *FILE;
76
77 open(FILE,'+<'.$self->{'_datfile'}) or carp $!;
78 binmode(FILE);
79 seek(FILE,$pos,0);
80 print FILE $val;
81 close(FILE);
82
83 return ($pos,length($val));
84 }
85
86 sub _addkey {
87 my ($self,$key,@pos_and_siz_pair) = @_;
88 local *FILE;
89
90 $self->{'_index'}->{$key} = \@pos_and_siz_pair;
91
92 open(FILE,'>>'.$self->{'_dirfile'}) or carp $!;
93 print FILE "'$key', ($pos_and_siz_pair[0], $pos_and_siz_pair[1])\n";
94 close(FILE);
95 }
96
97 sub TIEHASH {
98 my ($class,$file) = @_;
99 local *FILE;
100
101 my $hash = { };
102
103 $hash->{'_dirfile'} = $file.'.dir';
104 $hash->{'_datfile'} = $file.'.dat';
105 $hash->{'_bakfile'} = $file.'.bak';
106
107 $hash->{'_index'} = { };
108
109 sysopen(FILE,$hash->{'_datfile'},O_RDONLY | O_CREAT) or carp $!;
110 close(FILE);
111
112 my $self = bless($hash,$class);
113 $self->_update;
114
115 return $self;
116 }
117
118 sub EXISTS {
119 my ($self,$key) = @_;
120 return exists($self->{'_index'}->{$key});
121 }
122
123 sub FETCH {
124 my ($self,$key) = @_;
125 local *FILE;
126
127 my $pos = $self->{'_index'}->{$key}->[0];
128 my $siz = $self->{'_index'}->{$key}->[1];
129
130 open(FILE,'<'.$self->{'_datfile'}) or carp $!;
131 binmode(FILE);
132 seek(FILE,$pos,0);
133 read(FILE, my $dat, $siz);
134 close(FILE);
135
136 return $dat;
137 }
138
139 sub STORE {
140 my ($self,$key,$val) = @_;
141
142 if(not exists($self->{'_index'}->{$key})) {
143 $self->_addkey($key,$self->_addval($val));
144 }
145 else {
146 my $pos = $self->{'_index'}->{$key}->[0];
147 my $siz = $self->{'_index'}->{$key}->[1];
148
149 my $oldblocks = int(($siz + $_BLOCKSIZE -1) / $_BLOCKSIZE);
150 my $newblocks = int((length($val) + $_BLOCKSIZE -1) / $_BLOCKSIZE);
151
152 if($newblocks <= $oldblocks) {
153 my @pos_and_siz_pair = $self->_setval($pos,$val);
154 $self->{'_index'}->{$key} = \@pos_and_siz_pair;
155 }
156 else {
157 my @pos_and_siz_pair = $self->_addval($val);
158 $self->{'_index'}->{$key} = \@pos_and_siz_pair;
159 }
160 }
161 }
162
163 sub FIRSTKEY {
164 my $self = shift;
165 my $a = keys(%{$self->{'_index'}});
166 each %{$self->{'_index'}};
167 }
168
169 sub DELETE {
170 my ($self,$key) = @_;
171 delete($self->{'_index'}->{$key});
172 $self->_commit;
173 }
174
175 sub NEXTKEY {
176 my $self = shift;
177 each %{$self->{'_index'}};
178 }
179
180 sub UNTIE {
181 my $self = shift;
182 $self->_commit;
183
184 $self->{'_dirfile'} = undef;
185 $self->{'_datfile'} = undef;
186 $self->{'_bakfile'} = undef;
187 }
188
189 # it's true, baby ;-)
190
191 1;
192
193 # Documentation
194
195 =pod
196
197 =head1 NAME
198
199 DumbDBM_File - Portable DBM implementation
200
201 =head1 SYNOPSIS
202
203 use DumbDBM_File;
204
205 # Opening a database file called "homer.db"
206 # Creating it if necessary
207
208 my %db;
209 tie(%db,'DumbDBM_File','homer.db');
210
211 # Assigning some values
212
213 $db{'name'} = 'Homer';
214 $db{'wife'} = 'Marge';
215 $db{'child'} = 'Bart';
216 $db{'neighbor'} = 'Flanders';
217
218 # Print value of "name": Homer
219
220 print $db{'name'};
221
222 # Overwriting a value
223
224 $db{'child'} = 'Lisa';
225
226 # Remove a value
227 # The value remains in the database file, just the index entry gets removed,
228 # meaning you can't retrieve the value from the database file any more
229
230 delete($db{'neighbor'});
231
232 # Close the database file
233
234 untie %db;
235
236 =head1 DESCRIPTION
237
238 This is a Perl implementation of Python's C<dumbdbm> / C<dbm.dumb> module. It
239 provides a simple DBM style database written entirely in Perl, requiring no
240 external library.
241
242 Beware that this module is slow and should only be used as a last resort
243 fallback when no more robust module like L<DB_File> is available.
244
245 This Perl implementation is fully compatible to the original Python one.
246
247 =head1 FILES
248
249 Consider having a database called example, you have up to three files:
250
251 =over 2
252
253 =item example.dir
254
255 This is an index file containing information for retrieving the values out of
256 the database. It is a text file containing the key, the file offset and the
257 size of each value.
258
259 =item example.dir.bak
260
261 This file B<may> contain a backup of the index file.
262
263 =item example.dat
264
265 This is the database file containing the values separated by zero bytes.
266
267 =back
268
269 =head1 BUGS AND PROBLEMS
270
271 This module is a direct port of the Python module containing the same bugs and
272 problems.
273
274 - Seems to contain a bug when updating (this information was directly taken
275 from a comment in C<dumbdbm>'s source code)
276
277 - Free space is not reclaimed
278
279 - No concurrent access is supported (if two processes access the database, they
280 may mess up the index)
281
282 - This module always reads the whole index file and some updates the whole
283 index
284
285 - No read-only mode
286
287 =head1 COPYRIGHT
288
289 Copyright (c) 2019, Patrick Canterino, <patrick@patrick-canterino.de>
290
291 This Perl module is licensed under the terms of the 2-Clause BSD License, see
292 file F<LICENSE> or L<https://opensource.org/licenses/BSD-2-Clause> for details.
293
294 =head1 AUTHOR
295
296 DumbDBM_File was written by Patrick Canterino
297 L<patrick@patrick-canterino.de|mailto:patrick@patrick-canterino.de>.
298
299 L<https://www.patrick-canterino.de/>
300
301 If you wonder why I wrote this: I felt boring ;)
302
303 =cut
304
305 #
306 ### End ###

patrick-canterino.de